Strategic Thinking Assistant — Guided AI Conversations, Your Way
🚀 Stop getting walls of text. Start thinking strategically, step by step.
MindRoute is an Electron desktop chatbot that revolutionizes how you interact with LLMs. Instead of getting a single monolithic AI response, MindRoute generates multiple thinking paths — and lets you control exactly how deep the AI explores each one, step by step.
💡 Think of it like GPS for your thoughts — multiple routes to your answer, you pick the path and how far to go.
┌─────────────────────────────────────────────────────────────────┐
│ 🔵 You ask a question │
│ ↓ │
│ 🧠 AI generates 4 unique thinking approaches │
│ ↓ │
│ 🎯 You click a step → AI executes ONLY up to that point │
│ ↓ │
│ 📊 Get code blocks, diagrams, or structured analysis │
│ ↓ │
│ 🔄 New paths auto-generate based on the conversation │
│ ↓ │
│ 🔁 Keep exploring, switch directions, go deeper │
└─────────────────────────────────────────────────────────────────┘
| Step | What Happens |
|---|---|
| 1️⃣ | Type your question |
| 2️⃣ | Get 4 thinking approaches (Analytical, Creative, Practical, Visual Flowchart) |
| 3️⃣ | Click Step 1, Step 2, or Step 3 to control depth |
| 4️⃣ | Get structured responses with code blocks, Mermaid diagrams, or detailed analysis |
| 5️⃣ | New paths appear — keep exploring or change direction |
|
|
|
|
|
|
┌──────────────────────────────────────────────────────────────────┐
│ ● ● ● MindRoute ● Connected │
├────────────────────────────────────┬─────────────────────────────┤
│ │ ✦ THINKING PATHS │
│ 💬 Chat Messages │ │
│ │ ┌─ ANALYTICAL ──────────┐ │
│ Welcome to MindRoute... │ │ ① Break down question │ │
│ │ │ ② Examine components │ │
│ ┌──────────────────────────┐ │ │ ③ Synthesize insights │ │
│ │ 🧠 Following "Analytical"│ │ └───────────────────────┘ │
│ │ │ │ ┌─ CREATIVE ────────────┐ │
│ │ ```python │ │ │ ① Brainstorm ideas │ │
│ │ def solve(): │ │ │ ② Explore options │ │
│ │ return answer │ │ │ ③ Refine concepts │ │
│ │ ``` │ │ └───────────────────────┘ │
│ │ │ │ ┌─ PRACTICAL ───────────┐ │
│ │ ┌──────────────────┐ │ │ │ ① Implementation │ │
│ │ │ 📊 Mermaid Diagram│ │ │ │ ② Constraints │ │
│ │ │ A → B → C │ │ │ │ ③ Action steps │ │
│ │ └──────────────────┘ │ │ └───────────────────────┘ │
│ └──────────────────────────┘ │ ┌─ VISUAL FLOWCHART ────┐ │
│ │ │ ① Map process flow │ │
│ │ │ ② Decision points │ │
│ │ │ ③ Diagram connections │ │
│ │ └───────────────────────┘ │
├────────────────────────────────────┤ │
│ [What would you like to think...] │ ↻ Regenerate Approaches │
│ [Send] │ │
└────────────────────────────────────┴─────────────────────────────┘
MindRoute/
├── 🔧 main.js → Electron main process, Ollama SDK, optimized prompts
├── 🎨 index.html → Dark UI, code blocks, Mermaid rendering, sidebar layout
├── 📦 package.json → Dependencies and scripts
├── 📄 LICENSE → MIT License
├── 🚫 .gitignore → Node modules, build artifacts
└── 📖 readme.md → You are here!
| Requirement | Version |
|---|---|
| Node.js | v18+ |
| Ollama | Latest (for local mode) |
# 1️⃣ Clone the repo
git clone https://github.com/pandi-selvam/MindRoute.git
cd MindRoute
# 2️⃣ Install dependencies
npm install
# 3️⃣ Start the app
npm start# Auto-reload on changes
npm run devMindRoute connects to Ollama Cloud using gpt-oss:120b. To use your own API key, update main.js:
// main.js — initOllama()
ollama = new Ollama({
host: "https://ollama.com",
headers: {
Authorization: "Bearer YOUR_API_KEY_HERE",
},
});To switch to a local Ollama instance, update initOllama() in main.js:
ollama = new Ollama({
host: "http://localhost:11434",
});And change the model in all handlers:
model: "llama3.1:8b" // or gemma3:1b, mistral:7b, etc.| What | Where |
|---|---|
| Dark theme colors | index.html → :root CSS variables |
| Sidebar width | .paths-panel → width: 360px |
| Color accents | --purple, --blue, --cyan, --green vars |
| Window size | main.js → width: 1200, height: 800 |
| Shortcut | Action |
|---|---|
Enter |
📨 Send message |
Ctrl/Cmd + W |
❌ Close window |
Ctrl/Cmd + M |
➖ Minimize window |
F11 |
🔲 Toggle maximize |
Ctrl/Cmd + R |
🔄 Refresh thinking paths |
| Component | Technology |
|---|---|
| 🖥️ Desktop | Electron v27+ |
| 🤖 LLM Backend | Ollama (Cloud + Local) |
| 📡 API Client | Ollama SDK |
| 📊 Diagrams | Mermaid.js v10 |
| 🎨 Frontend | Vanilla HTML / CSS / JS (dark glassmorphism) |
| 🔧 Dev Tools | concurrently, nodemon |
| Domain | How MindRoute Helps |
|---|---|
| 💻 Development | Debug with controllable depth, get code blocks with copy button, explore architecture with flowcharts |
| 📚 Research | Break complex topics into structured, step-by-step exploration with visual diagrams |
| 🎯 Decision Making | Evaluate options through analytical, creative, and practical lenses |
| 🔍 Problem Solving | Systematic root-cause investigation with Mermaid flowcharts mapping the process |
| 📝 Writing & Planning | Outline strategies with incremental, guided thinking and visual flow mapping |
- 📊 Mermaid.js flowchart diagrams
- 💻 Syntax-highlighted code blocks with copy
- 🎨 Dark glassmorphism sidebar layout
- ☁️ Ollama Cloud integration
- 🎛️ Multi-model support — different models per path
- 📋 Custom thinking templates — save & reuse approaches
- 💾 Conversation export — save sessions as documents
- ⚡ Streaming responses — tokens as they arrive
- 🔌 Plugin system — extend with custom tools
Contributions are welcome!
# 1. Fork the repo
# 2. Create your branch
git checkout -b feature/awesome-feature
# 3. Commit your changes
git commit -m "Add awesome feature"
# 4. Push and open a PR
git push origin feature/awesome-featureMIT License — see LICENSE for details.
- 🦙 Ollama — Making LLM deployment accessible
- ⚡ Electron — Cross-platform desktop framework
- 📊 Mermaid.js — Diagrams and flowcharts from text
- 💭 Inspired by consulting methodologies and structured problem-solving
Built with 💜 by Pandi Selvam P
🧠 Think smarter. Explore deeper. Control your AI conversations.